"

 

VESUAL ACUITY AS AFFECTED BY
ADJACER‘F EORDERS N A TARGET

Thesis for the Degree of DH. D.
MECBIGEN STATE UNEVERSITY

J. Yves Lorrie
1965

 

THESES

LIBRARY—L1

Michigan State
University

 
     

This is to certify that the

thesis entitled

Visual Acuity as Affected by

Adjacent Borders in a Target
presented by

Jean Ives Lortie

has been accepted towards fulfillment
of the requirements for

PhoDo degree in PSXCh010gy

WJWSZ‘ 0’“ ’fi L

. K
Major professor \

17d?

Date MM 3

0-169

ABSTRACT

VISUAL ACUITY AS AFFECTED BY
ADJACENT BORDERS IN A TARGET

by J. Yves Lortie

This investigation examines the effect of adjacent
borders upon detection of a fine line.

A review of the literature on contour processes indi-
cates that borders near one another exert a mutually
depressive influence. This influence was shown to extend
to distances as far as four degrees on the retina. If in
a set-up traditionally used for studying visual acuity two
dark bars are introduced at different distances from a fine
line to be detected, one would expect visual acuity to be
impaired when the bars are close to the line. As the dis-
tance is increased, the depressing effect of the bars
should progressively decrease. Moreover, such a depressive
effect, if present, can be better understood by performing
temporal manipulations of borders. Thus, varying the order
of presentation of the bars and line should reveal signifi-
cant interactions between contours. It should also throw

some light on the relationship between contour processes

and visual acuity.

J. Yves Lortie

It was the purpose of this author to examine those
points, or more generally to analyze some of the impli-
cations of relating visual acuity and contour processes.

Two trained observers participated in the experiment.
Six situations were investigated in which several spatial
and temporal manipulations of borders were performed, such
as length of bars and line, distance of bars from the line,
order of presentation of bars and line, and shifting of
bars.

The results indicate that visual acuity is affected by
the presence of borders in the vicinity: (1) Distance be-
tween the bars turned out to be a significant factor.

When these bars were near the line, more time was required
for its detection than when they were far out. At an inter-
mediary distance, however, they had a facilitatory effect
upon visual acuity. (2) The length of bars, when near the
line, counterbalanced the facilitatory effect of an increase
in length of this one. (3) Removal of the bars at the same
time as the line was projected had a strong depressive
effect upon visual acuity. (4) A shifting of the bars away
from the line also required more time for its detection.

(5) A shifting of the bars toward the line was facilitatory
in the case of one observer, and inhibitory in the case of
the second one. (6) Equally, for one observer simultaneous

presentation of the bars and line had a facilitatory

J. Yves Lortie

influence upon visual acuity while the effect was inhibitory

for the other.

The results were discussed in terms of contour forma—
tion. They suggest that any factor which impairs visual
acuity does so by interfering with this neural process.
Among the factors found to play a significant role are the
presence of other contours in vicinity, their order of

formation and their destruction.

Approved: gfiQNA—BS;)~wQL

CommitteeSChaifman

Date: YMM:21gcwb2f

VISUAL ACUITY AS AFFECTED BY

ADJACENT BORDERS IN A TARGET

BY

.— ‘t' '
f 1

J3 Yves Lortie

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Psychology

1965

ACKNOWLEDGMENT S

The author is indebted to Dr. S. Howard Bartley
for his advice and supervision. His many helpful sug-
gestions led to the realization of the present research.
He is also grateful to Dr. H. Richard for his encourage-
ment and for the provision of facilities in the Depart-

ment of Psychology of Laval University, Quebec.

ii

TABLE OF CONTENTS
Page
INTRODUCTION . . . . . . . . . . . . . . . . . . 1
METHOD . . . . . . . . . . . . . . . . . . . . . 25

Subjects . . . . . . . . . . . . . . . . . 25
Apparatus and Material . . . . . . . . . . 26
Preparation of Subjects. . . . . . . . . . 31
Procedure. . . . . . . . . . . . . . . . . 55

RESULTS . . . . . . . . . . . . . . . . . . . . 59
DISCUSSION . . . . . . . . . . . . . . . . . . . 65
SUMMARY . . . . . . . . . . . . . . . . . . . . 82

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . 85

iii

LIST OF TABLES

TABLE Page
1. Raw data on threshold duration obtained in
some of the situations—-for the two ob-
servers. Photic intensity .055 c/ft2 . . . 59
2. Raw data found to be less stable. Photic
intensity .055 c/ft2. Observer J.V. . . . 40

iv

LIST OF FIGURES

FIGURE

1.

2.

10.

11.

12.

Targets 1 F and 1 E used in fore and ex-
posure periods respectively. . . . . . . .
The six experimental situations examined

under three different levels of photic in-
tensity. . . .

Visual acuity as dependent upon the dis-
tance of adjacent borders. . . . . . . . .
Effects of diverses manipulations of ad-

jacent borders upon detection of the line.
Data for J.Y.L. . .

The influence of diverse manipulations of

adjacent borders upon visual acuity. Data
for J.V. . . . . . .

Comparisons of sets of data collected at a
six month interval . . .

Visual acuity as affected by a shifting of
position of adjacent borders in fore and
exposure periods . . . . . . . . . . . . .
The relation between bars length and

visual acuity for two different distances
of the bars. . . . . .

Visual acuity as dependent upon bar length
and distance between the bars. . . . . . .
Comparison of duration thresholds for five
of the experimental situations. Data for
J.Y.L. O O O O O O O O O I O O O I O O C 0
Mean detection thresholds for five of the
situations, with J.V. as the observer. . .

Gradient-wise projection of a dark bar
upon the retina. . . . . . .

Page

29

34

41

44

45

48

51

54

56

59

62

69

INTRODUCTION

Visual acuity is a spatial discrimination involving
a simultaneous comparison of different areas of a target.

It is defined as the reciprocal of the minimum visual angle
subtended by some relevant element of the target, measured
in minutes of arc. In order for visual acuity to take
place, there must be a detectable disparity or lack of
uniformity in the different portions of the target; in other
words, borders must exist in the target itself.

A recent review (Westheimer, 1965) indicates that many
different data have been collected concerning visual acuity.
Equally, various explanations have been offered: the most
important of these have been critically reviewed by Senders
(1948) and Falk (1956). The discussion has been brought
up to date by Boynton (1962)"

However, the detection of a fine line, a grating, or
any other figure, implies the formation of contours (Bartley,

J.

1941, p. 354). This neural mechanism has not been given

all the attention it deserved in psychophysical as well

 

J'Because he is dealing with three different types of
phenomena: physical, neurophysiological, and perceptual,
the author, following Bartley (1958 a, p. 150), differen—
tiates between borders, contours, and edges. The abrupt
changes in illumination of the target are called borders,
the corresponding neural processes are named contour forma—
tion or contour processes and the perceptual end-results
are the edges.

1

as in neurophysiological researches. For example, one of
the major findings of the studies devoted to it is that
borders or abrupt changes of luminance in a target, at
certain distances one from the other, interact in the
formation of their reSpective contours, thus changing the
perceptual outcome (Fry & Bartley, 1955). Any condition
affecting this process, as for example, the depressing in-
fluence of borders in near vicinity, should change visual
acuity. Very few studies, however, have approached the
problem in this way. It was the purpose of the present
author to examine the evidence offered about this process,
to study its relation to visual acuity, and to perform
further manipulations of borders in order to observe their
effects upon detection of a fine line.

In this chapter some of the studies on contour forma-
tion are reviewed and generalizations to which they led are
examined. A review of these generalizations in relation
to visual acuity follows. Finally, Specific questions are

formulated which it is the purpose of the author to study.

Contour Processes

Definition
When a target with abrupt borders is projected upon
the retina, the information which reaches the cortex through

the optic tract can be conceived of as being of two types:

(1) a longitudinal propagation of activity along the dif-
ferent channels of the pathway; (2) a lateral or cross-
sectional type of activity which takes place in the retina
and across the channels (Nelson, Bartley & Wise, 1963).
This lateral type of discharge has been called contour forma—
tion (Bartley, 1941, p. 229) or contour processes (Bartley,
1958 a, p. 150). Thus, contours are neural processes under—
lying Spatial discriminations in which edges are involved.

Contour processes have been posited in order to deal

with these facts:

(1) the retinal image of a sharp border is blurred;

(2) brightness discrimination is affected by the
distance between borders in a target, and by the
timing of successive presentations of borders;

(3) the perceptual or phenomenological properties
of a target differ depending on the degree of

illumination used.

Blur of the Retinal Image

In prOper conditions, abrupt borders in any portion
of a target lead to the perception of steep, sharp edges.
However, the corresponding image formed on the retina,
besides being upside down, is more or less blurred, never
as clear-cut as the borders themselves. In other words,
the image on the retina is not an exact replicate of the

physical object impinging on it. As the edge is seen as

clear-cut, some neural mechanism, conceived in terms of
enhancement and depression of lateral activity, is needed
to resharpen what has become blurred.

The gradient-wise distribution of retinal illumination
results mainly from diffraction effects and different aber-
rations of the refractive power of the eye. These two types
of defects vary with the size of the pupil. Thus as the
pupillary diameter decreases, diffraction effects become
larger (Westheimer, 1963). With large pupils diffraction
becomes negligible but, on the other hand, chromatic aberra-
tion comes in: it depends on different wavelengths being
refracted by different amounts, and this creates indistinct
color effects along the edges of the perceived image.

Another type of aberration found in optical systems
is spherical aberration: the rays refracted through the
outer portion of the lens are brought to a focus nearer that
lens than those from the center. This normally would pro-
duce a blurred and distorted image. This factor, however,
is negligible in the eye: it is compensated to a high
degree by the peripheral flattening of the cornea (Bartley,
1960, p. 202; Westheimer, 1963), and also by the size of the
pupillary aperature and the structure and curvature of the
lens (Bartley, 1960, p. 202).

Other characteristics of the eye also affect the amount
of blur of the retinal image: one of these is the accommo—

dative power of the eye. Poor accommodation increases the

amount of blur by bringing the image out of focus (Westheimer
& Campbell, 1962). Moreover, it has been found that, during
fixation, accommodation is constantly changing (Campbell &
Westheimer, 1960; Arnulf & Dupuy, 1961).

Finally, physiological nystagmus would be assumed to
decrease the steepness of the gradient: the rapid oscilla-
tions of fixation successively present adjacent areas of the
target to the same retinal receptors and this should enhance
the blur of the formed image. The author, though, did not
find any direct evidence supporting this view.

Distance Between Borders in Brightness
Discrimination

Suppose a small disk is surrounded by a larger ring
and that the luminance of the disk is increased till the ob-
server perceives a difference in brightness between the disk
and its annular surround. It has been found (Blachowski, 1915)
that the threshold for the detection of the disk decreases as
the area of the surround is increased. Blackowski explained
his results in terms of Spatial summation. Fry & Bartley
(1935), however, have demonstrated that the decrease in thres-
hold detection when the area is enlarged is due to the fact
that in these conditions the outer borders of the ring are
farther away from the borders of the disk and they interfere
less with the formation of the appropriate contour.

In their experiment they used a disk with a diameter of

5/8 inch and a large 6 inch ring surrounding it. Also a dark

circular band, the diameter of which could be increased or
decreased, was superimposed on the ring. By decreasing the
diameter of the dark band, the distance between its inner
border and the border of the disk was decreased, and the
opposite was observed when the experimenter was increasing
the diameter of the band. The area of the ring, however,
remained constant. It was found that when the distance
between the borders of the disk and the inner border of the
band was 4 degrees or less the threshold of detection of the
disk was higher; it increased in proportion as the distance
between those borders was decreased. The authors concluded
that already formed contours exert a depressive influence
upon the formation of other contours in their immediate sur-
round.

Other manipulations by the Same authors allowed them
to reach a second conclusion, that is: contours parallel
to the activating contour are depressed while those perpen—
dicular or at right angles to it tend to be facilitated.
Thirdly, existing contours prevent activity from Spreading
in the visual system; a third contour interposed between two
activating contours will block the activity between these
two.
Timing of Successive Presentations of
Borders

The work of Fry & Bartley, reviewed above, has indicated

the effect of Spatial distribution of events upon contour

formation. Timing of events is a second important variable.
Helson & Fehrer (1952) presented targets of different shapes
tachistoscopically. With a very brief exposure, these were

seen as "dim patches of light." As the exposure was length—
ened, definite forms with sharp edges were recognized.

In an extensive experiment, Werner (1955) used many
different targets and manipulated Spatial as well as temporal
variables. For example, in a part of the experiment he had
a small illuminated disk alternate with an illuminated ring
the inner border of which coincided with the outer border of
the disk.

When the disk was briefly presented first and was
followed after about 150 msec. by the ring, the disk was not
seen. The outcome was the same, when the figures were dark
and the ground illuminated. With a slower rate of succession,
the disk was seen and then the ring. Werner concluded that
with a fast rate of succession, contours do not have time to
form and consequently the figure is not perceived as such.
Too short a duration of presentation depresses or destroys
the formation of contours and it affects the appreciation
of the brightness that a surface would otherwise have.

Other shapes, where the space between two wide dark bars
was occupied by several thin lines parallel or perpentidular
to these bars yielded results sensibly Similar to those of
Fry & Bartley for the parallel lines: these wereystrongly

affected by the previous presentation of the wide bars.

The lines perpendicular to the bars were depressed also, but
to a lesser degree. Fry & Bartley had found that borders at
right angles tended to facilitate contour formation. Fry &
Bartley, however, were manipulating Space or distance be-
tween borders, while Werner manipulated duration of presenta-
tion in that part of his experiment.

Role of Luminance Upon the Edge

Properties of a Target

The intensity of illumination of the different portions
of a target is another variable which has been shown to con-
siderably alter the formation of contours. Suppose (Bartley,
1941, p. 6) a stimulus consisting of a small illuminated disk
(spot of illumination) in a dark field. When the intensity
is very low, the observer sees an indefinite Spot of light,
not well differentiated from the surround, which seems to
wander about and may even disappear for brief moments.
Phenomenologically the target is not seen as a disk, but as
an indifferentiated Spot. If the intensity of illumination
is increased sufficiently, then the result is quite different:
what emerges is a disk with Sharp edges.

An eXperiment which demonstrates the role of illumi-
nation upon the edge properties of a disk has been performed
by Bartley (1956). He used a target arrangement consisting
of a disk surrounded by a ring the inner border of which
coincided with the outer border of the disk. The disk portion

of the target was alternately light or dark at each of several

Slow rates, while the ring portion was kept constant in
illumination. Physically the duration of the light and dark
phases of the disk was the same; however, the intensity of
illumination of the ring determined the perceptual outlook:
when the intensity of the ring was kept above the mean value
of the two phases of the disk, the dark phase of the disk
became predominant. When the intensity of the ring was re-
duced to a value below the mean, it was the light phase which
precominated. The predominant phase possessed sharp edges
and seemed to occupy most of the cycle, while the diminished

phase had no definite edge but was seen as a mere shadow.

Contours in Relation ot Other Phenomena

Contour processes are involved not only in brightness
discrimination, but also in many other phenomena: Bartley
(1941) presents some evidence to the fact that contour
formation has a decisive role in Fechner's paradox, in after-
images and in visual acuity. Osgood (1955, p. 252) considers
contour formation as being fundamental to all perceptual
activity.

The relation between contour processes and visual
acuity will be examined after the following section on neuro-

physiological researches.
Neurophysiological Studies on Contours

Although several psychophysical researches Show the

legitimacy and importance of contour processes, these have

10

not been much investigated as such by neurophysiological
techniques. This may be due to the complexity of the
processes themselves: for example, it seems that in order

to study them adequately large areas of the retina or many
different channels of the optic tract would have to be
covered simultaneously. In recent years microelectrode
studies have tended to explore in a piecemeal fashion very
restricted regions of the pathway, while macroelectrode
techniques did not systematically investigate those processes.

However, there has been some interesting work on what
can be considered as part-effects of contour formation, that
is on inhibitory mechanisms. For example, Hartline and
collaborators (Ratliff, 1961), working with the lateral eye
of the horseshoe crab, Limulus, and using diffuse illumination
as a stimulus, reached the conclusion that the detection and
enhancement of edges is explainable in terms of lateral in-
hibitory interaction among retinal elements.

While Hartline and collaborators.were working at the
retinal level, Jung and co-workers (1961) investigated the
visual cortex, recording neuronal discharges in the primary
visual area of the cat. The stimulus used was diffuse illumi—
nation. For binocular stimuli, they proposed to classify the
response patterns of the visual cortex into five neuronal
types, named A to E, thus expanding the classification of
the retinal elements which comprises on, off, and on—off

neurons. As no physical borders were used, not much could

11

be reported related to perception of edges. But, in 1958.
Baumgartner, in the same laboratory, began to use what he
calls "patterned light with white-black contrast," or targets
with abrupt borders (a grid of light and dark bars). In one
of their experiments, he and Hakas (Jung, 1961) exposed one
of the light bars, subtending a visual angle of 5 degrees

41 minutes, and moved it by steps across the receptive fields
of the cortical neurons. As expected, the response of these
neurons was different from what is observed when there is

no borders (diffuse illumination). When the illumination is
on, a maximum of discharges is observed in the cortical B
neurons if their receptive field is stimulated by the portion
of the light bar at the margin of the dark bar (at the border).
When the receptive field of these same neurons is presented
with the dark ground, there is a minimum of discharges at
light-on and also a reversal of their responses to onset the
termination of illumination, that is when the illumination is
on, the discharge of the B neurons is inhibited; it is
activated when the illumination is off. The D neurons behave
in the opposite way; they are activated when-the illumination
is on, and depressed when it is off, if their receptive fields
are presented with the dark part of the target. When the
light part is used, the off-reSponse of these D neurons is
maximal. To explain this reciprocal activation and inhibition
of these two types of antagonistic cortical neurons and also

the increased frequency of discharges at the border between

12

the light and dark bars, they utilized the reciprocal and
lateral inhibition schema of Jung, that is there is a recipro-
cal inhibition of antagonistic neurons (B and D) in the same
receptive field, and a lateral inhibition of synergic neurons
in the surrounding regions.

Another interesting finding of Baumgartner is that the
neuronal discharges corresponding to the borders between
light and dark bars vary with the width of the bars, a fact
in agreement with psychophysical observations to be reviewed
later.

From what has been done till now, one can easily con—
clude that physiological studies are still far from providing
the information necessary to clarify the mechanism of contour
formation. The conclusions reached by Hartline when applied
to the human or even the vertebrate retina at a lower level
(the cat, for example) are very limited, at least for the
reason that the compound eye of the Limulus is structurally
and functionally very different from the vertebrate eye.
Hartline and co-workers agree that in the vertebrate retina
interaction would be more complex and would comprise excita-
tory as well as inhibitory influences (Ratliff, 1961, p. 200).
Also it is to be recalled that Hartline was recording neural
impulses in one active nerve fiber at a time, thus reaching
conclusions very limited in scope.

Baumgartner worked with a vertebrate organ, the cat

eye, and also used borders in his target. Second, he took

I

15

care of comparing his physiological findings (neuronal dis-
charges in cats) with corresponding visual perceptions in
man, thus coming closer to bringing the elements which would ‘
allow a thorough understanding of the process of contour
formation. However, here also one finds serious limitations:
for example, his conclusions are valid only when eye move-
ments are excluded. Optokinetic nystagmus seems to be an
important factor in the blurring of the retinal image as it
was mentioned previously. Besides, here again the work done
was very analytical, in a piecemeal fashion, the authors
recording from one receptive field at a time, while contour
formation seems to work over very large areas of the retina,
as the psychophysical studies of Bartley have Shown. In
short, much more researches are needed before one can trans-
late into clear physiological terms the processes of contour
formation. Therefore, it is not the purpose of this author
to work out such a mechanism. Attempts have been made else-
where (Osgood, 1955, p. 229; Milner, 1958), but it is the
author's Opinion that they are premature. For the time being,
more work should be performed on the psychophysical as well

as on the physiological levels: physiological researches

will offer direct evidence on contour processes, while psycho-
physical ones will bring in new facts and data which eventually
one should be able to explain if he wants the processes to

be of more than very restricted validity.

14
Contours Processes and Visual Acuity

While in studies involving brightness discrimination
one was working with extended areas, in visual acuity the
area becomes minimal, at least along one dimension, and
borders are much nearer one another. This effect of borders

upon one another Should, therefore, become more prominent.

Manipulations of Borders in Visual Acuity

The distance between borders, in a visual acuity target,
can be looked upon in several different ways: for example,
in a target made up of a Single line there is the distance
between the two longitudinal and parallel borders of the
line, the distance between the borders at the two extremities
of the line, and also the distance between the line and the
borders formed by the frame of the whole target. When the
target consists of two parallel lines, there is, besides the
factors just mentioned, the interspace between the two lines,
In a grating, there are Several interSpaces. With a Landolt
ring, the relation between borders is more complex, while
with letters the complexity of the relations is very variable,

All of these shapes, and others, have been used in
studying visual acuity. For example, disks (Ogle, 1961) as
well as fine wires or lines (Fry & Cobb, 1955; Hecht & Mintz,
1959) have been used for studying the minimum visible. For

the minimum separable the following configurations have been

utilized: pairs of parallel bars (Fry & Cobb, 1955; Wilcox,

15

1952), Landolt ring (Shlaer, 1957), grating (Shlaer, 1957;
Graham & Cook, 1957), two points (Oliva & Aguilar, 1957),
and vernier adjustments.(Baker, 1949; Leibowitz, 1955).

Among these, a very few are relevant to the problem
with which the present author is concerned, those where
borders in simple relation to one another are implied.

Single line. The minimum width for a single dark line
to be visible was Shown by Hecht & Mintz (1959) to be nearly
0.5 second at the highest illumination they used (50 milli-
lamberts) and with binocular vision. The target was made up
of a circular ground, an opal glass illuminated from behind
and measuring 2 feet in diameter. The lines were made up
of wires varying in thickness. The subject, in a chair on
rollers, moved toward the target till he could ascertain the
position of the line in the target. Hecht & Mintz explained
its detection by the fact that the illumination on one row
of cones is just perceptibly less than on the rows on either
side of it. So, even if the retinal image of a fine line is
fuzzy, it is not perceived as such but it is seen as a clear
line because it would stimulate one row of cones less than
the others adjacent to it. They maintained that no central
mechanism of contour formation is necessary to convert the
gradual distribution of illumination on the retina into a
Sharp line at the perceptual level.

Their eXplanation runs into serious difficulties. These

have been covered by Senders (1948),and Falk (1956). Let‘s

16

mention only that according to Hecht & Mintz visual acuity
would depend only on intensity of illumination. As will be
seen, other factors are also implied, for example the length
of the line.

In the previous experiment, Hecht & Mintz always used
wires of the same length. In another experiment, Hecht,
Ross & Mueller (1947) varied the length of the fine wires.
Again they found 0.45 sec. as the minimum width of a line
to be seen against a bright sky. They found, however, that
in order to get this result, the minimum length had to sub-
tend an angle of about one degree at the eye. Below that
value the threshold raised rapidly.

The facilitating effect of length of line was confirmed
by Ogilvie & Taylor later on (1959).

The effect of the width of a light bar on visual
acuity has been investigated by Fry & Cobb (1955). Using
bars 50 minutes long they determined the threshold values
for a number of different widths, beginning with a bar whose
width was 5 minutes 12 seconds and then using bars of lesser
width till they obtained one in which width and intensity
were reciprocal. They were able to demonstrate that the
retinal distribution of illumination is Gaussian in character;
therefore, it is the intensity at the center of the retinal
image of the bars which determines visual acuity. With a
very narrow bar the intensity at the center may be subthres—

hold. Increasing the width of the bar increases the intensity

17

at the center, so that the line may then become visible.
However, one reaches a point where this relationship of
width and intensity does not hold any longer; for example,
any bar wider than 4 minutes 2 seconds, in the conditions of
the experiment, did not reduce the threshold of detectability.
It is to be noted that in this direct method of estimating
the blur of the retinal image, nystagmus or the rapid oscilla—
tory movements of the eye as well as diffraction and the
chromatic and Spherical aberrations were taken into account.
That the retinal gradient of illumunation is Gaussian
or approximately so has been confirmed by Westheimer and
Campbell (1962). In their experiment they used a streak light
filament. They estimated the Spread of the corresponding
retinal illumination by determining the distribution of in-
tensity in the aerial image formed by reflection of the
filament at the fundus and subsequent reverse passage through
the eye.

Parallel lines. When one is using two parallel bars

 

visual acuity has been shown to depend not only on the inter—
Space between these bars but to a certain point on the width
of the bars themselves. For example, Fry & Cobb (1955) in
the same series of experiments as described above, used two
pairs of light bars measuring 55 minutes 20 seconds in length.
One set of bars were 16 minutes 40 seconds in width while

the second set, the narrow bars, were 2 minutes 48 seconds.

Their results indicated that with an increase in the

18

luminosity of the bars from 0 to 5 foot-candles visual acuity.
in the case of the wide bars, always increased, rapidly at
first and then more slowly. In the case of the narrow bars,
visual acuity improved markedly till the luminosity reached
slightly more than one foot-candle, and then it deteriorated.
To explain their findings the authors utilized the principle
established by Fry & Bartley, namely that borders near one
another interfere in contour formation. With narrow bars,
parallel borders are near one another, the contour processes
corresponding to each one of these borders interfere and
depress each other thus elevating the threshold for detection
of the interspace. With wider bars, the borders are farther
apart and their influence is much less marked.

The importance of bar width in determining visual
acuity was further studied by Kravkov (1958). Wilcox (1952),
studying the effect of illumination upon visual acuity, had
found that with dark bars on a light ground detection of
the interSpace between the two bars increased with an in-
crease in retinal illumination, but that with light bars on
a dark ground, visual acuity first improved and then de-
teriorated. Furthermore, he observed that dark bars tended
to be subjectively widened at low intensity, and that this
apparent enlargement was decreasing with an increase in
intensity, the bars subjectively appearing to Shrink.
Kravkov showed that the conclusions of Wilcox are valid for

narrow bars only, not for wide bars. He explained this

19

phenomenon of shrinking, called negative irradiation by
Wilcox, in terms of the findings of Fry & Bartley on the
depressing influence of borders parallel to one another on
contour formation. In brief, with narrow bars two factors
are at work, one photic intensity, favoring visual acuity,
the other, borders near one another, impoverishing it by
creating a shifting in contour processes toward one another.
With an,increase in photic intensity, the increased steepness
of the blur gradient would overcome the depressing effect of
the other factor and an improved visual acuity would result.
With broad bars, photic intensity becomes the main factor

and no shifting is observed.
The Present Problem

The studies reviewed above, especially those by Kravkov,
Fry & Cobb, indicate that borders near one another affect
each other in determining visual acuity. Moreover, the work
of Fry & Bartley on brightness detection had shown that this
depressing influence may extend to distances as far as 4 de-
grees on the retina. If a whole target subtends a visual
angle of only a degree or so, which is very often the case
in experimental set-ups where visual acuity is investigated,
then the borders formed by the frame of the target should
exert some influence on the perception of the line or lines.
For example, the background field in Wilcox's study subtended

an angle of 1 degree 10 minutes in width and 20 minutes in

20

height. According to the generalizations of Fry & Bartley,
these borders of the field which are parallel to the two
lines should affect their detection. Second, those borders
perpendicular to the lines should exert a facilitatory in-
fluence. As the lines are relatively short, 20 minutes in
length, this influence should be serious. But these vari-
ables have not been dealt with in Wilcox's study.

The same problem is encountered in any set—up wherein
relatively small background fields are used.

It seems then that systematically examining the effects
of bars presented at varying distances from a fine line
would be justified: besides throwing new light on the pro—
cesses investigated, such a study would allow one to test
the validity of using a background field of restricted di—
mensions and at the same time ignoring its effects when
theorizing about visual acuity.

In fact, one would expect visual acuity to be impaired
if these borders are in close vicinity of the line. As the
distance is increased, the depressing effect of the bars
should progressively decrease. As simple as this may seem,
the present author does not know of any study trying to
investigate it. One study came close, though: Flom and co—
workers (Flom, Weymouth & Kahneman, 1965) examined the effect
of four dark bars placed tangential to a Landolt ring and
at varying distances from it, upon detection of the position

of the gap in the ring. The maximum interaction was found at

21

approximately five times the gap width. However, in that
study the relation between borders is very complex; it is
far from being the ideal situation for unraveling the func-
tioning of contour processes.

Another aSpect of the problem refers to temporal
manipulations of borders. As contour processes take time to
complete themselves, it appears that varying the order of
presentation of different portions of the target would be
essential for one to better understand the relationship
between contour formation and visual acuity. For example,
the bar borders: the bars borders may be presented before
the line to be detected, or Simultaneously with it, or they
may be removed at the exact moment the line is projected,
etc.: all these manipulations should reveal significant
interactions between borders. To the author‘s knowledge,
such manipulations have not been tried. Traditionally all
the elements of a target have been presented at the same
time. In a typical experiment the target is presented for
unlimited time and the observer indicates if he can detect
or not the relevant portion (Hecht & Mintz, 1959). Or, while
the observer is looking at the target, the distance between
bars or lines may be varied by the experimenter till these
are viewed as separate (Wilcox, 1952). In other experiments,
the observer has a limited time for looking at the target,
but again all the relevant portions of the target are pre-

sented together for a certain duration (Niven & Brown, 1944).

22

Rare are the exceptions to that procedure: for example,
Granit & Harper (1950) used the c f f method, but their
purpose was to study the relation between spatial summation
and visual acuity. Equally, Bartley and collaborators
(Bartley, Nelson & Soules, 1965) utilized intermittent
illumination; this was done, however, in order to investigate
the effect of brightness enhancement upon visual acuity.
Thus, in these two types of studies temporal manipulations
of different portions of a target were effected but they were
not of the kind suggested above.

Third, another generalization of Fry & Bartley Should
lead to interesting studies in visual acuity. They have
shown that already formed contours prevent activity from
Spreading in the whole retina. Therefore, in specific con-
ditions to be determined, already formed contours should
improve visual acuity. As far as the present author knows,
this generalization has not been investigated.

The purpose of the present study was to examine those
factors, or more generally to analyze some of the impli-
cations of relating visual acuity and contour processes.

If distance and timing of borders can be shown to affect
visual acuity, a new type of evidence will be brought forth
for the processes involved. In order to determine whether

this is so, the effects of diverse Spatial and temporal

manipulations of borders were analyzed:

(1)

(2)

25
manipulations of length of borders and distance
between them;
temporal manipulations of different portions of

the target.

Within this general framework, Six Specific questions

were studied:

(1)

(2)

(5)

(4)

(5)

Would a decrease in the distance between two dark
bars retard the detection of a fine line appear-
ing in the interSpace between those bars?
Similarly, would a dark foreperiod followed by a
simultaneous presentation of the bars and line,
hinder visual acuity?

To what extent would a lighted foreperiod without
any border in the target and followed by a simul-
taneous projection of the bars and line, affect
the detection of this same line?

Would a foreperiod with two bars which vanish when
the line appears, create much interference upon
seeing the line?

Would an instantaneous shifting of the positions
occupied by the bars while the line is projected,
increase the duration required for detecting the
line? If so, would the effect be different when
the shifting is toward the line than when it is

away from it?

24

(6) Finally, would longer bars cancel out the facili-
tatory influence of an increase in the length of
the line upon visual acuity?

The general plan of this research involved testing

two trained observers for visual acuity under three different
levels of photic intensity, in situations wherein Spatial

and temporal manipulations of targets were performed.

METHOD

Subjects

Two subjects were used throughout the whole experi-
ment: J.V., a graduate student in Education (male, 55 years
old) and the author himself (male, 55 years old).

J.Y.L. had his vision corrected for my0pia. As for
J.V., he declared, at the beginning of the experiment that
he had normal vision. However, a few months after the second
part of the experiment was completed, that is in April 1965,
he had his eyes examined and a Slight degree of hypermetropia
(+1.50 diopter) was found. The implications of this re-
fraction defect are discussed in the prOper section of this
report.

J.V. had been trained thoroughly in May 1964: he then
spent more than fifty hours as experimenter or subject in a
first attempt at studying contour formation. The data col—
lected at that time had to be rejected when the whole pro—
cedure was modified later on. However, the experience
acquired by both J.V. and J.Y.L. led to more reliable data
in the second attempt. AS Wilcox (1952) mentions, visual
acuity for fine lines varies with practice.

J.V. though well practiced, was not aware of the
hypotheses to be tested and he knew vaguely about the process

of contour formation.

25

26

Nelson, Bartley and De Hardt (1960) have shown that
data obtained from well-trained subjects are alike in pattern
and that they are close to the mean found with more naive
subjects. In order to check that point, the author, at first.
expected to have two naive subjects besides the two trained
ones. It was not possible, at this period of the year, to
find students who would spend a month or so in the laboratory.
Only one subject was found (D.V., male 55 years old, a rela-
tive of J.V.). He was trained during two days and then data
were collected. However, heirepeatedly failed the test tar-
gets (to be described later) and his data had to be rejected

entirely.

Apparatus and Material

The apparatus was the Gerbrands Tachistoscope, with .
two illuminated G backs and the associated Dual Timer.
Because the illumination of the G backs was not uniform ex-
cept for a very small area, the apparatus was slightly modi—
fied in the following ways the two neon bulbs, in each G
back, instead of being about one inch behind the milk glass,
as in the original set-up, have been moved approximately six
inches away from the glass; a wooden frame, painted white
inside, completed the arrangement. As a result the illumie
nation became uniform to an acceptable degree over the entire
area allowed by the tachistoscope (7%" x 7%").

Second, the small rubber head-set on the Tachistosc0pe

was replaced by a 14" X 20" wooden black board with three

27

sides protruding 5" in order to block any photic radiation
coming from other sources than the one to look at. An ad-
justable chin-head rest (American Optical Company) was used
which was much more reliable and comfortable than the
original set-up besides allowing one to wear glasses. When
properly installed, the observer has his right eye 24" from
the source and viewed the different targets through a 7/8"
peephole located 1%" from his eyel: this set-up allowed an
angle of vision of 55 degrees. A piece of black paper pro-
truding 1" was put to‘the left of the peephole in order to
impede any illumination from reaching the left eye. With
the right eye closed, only with full illumination on was it
possible for the observer's left eye to detect a certain
amount of brightness, and it was so low it was assumed not
to interfere with the behavior of the right eye during the
experiment prOper.

The illumination was varied by means of Cinemoid Grey
Filters (No. 60) introduced immediately behind the peephole.

Calibration of the Dual Timer was accomplished with a
K M 1 Short Timer (which could measure from 0.005 msec. to

10 sec. with an accuracy of 2%). Afterwards, the main source

 

1No artificial pupil was used for fear that it would
not be in line with the natural one. As the purpose of the
author was not to measure the highest visual acuity attain-
able, in which condition the pupil Size would become import—
ant, but to study the effetts of borders upon contour forma—
tion in visual acuity, the pupil diameter becomes a less
important variable.

28

of error resulted from trying to duplicate the different
settings: a very Slight variation in the successive settings
for a specific value yielded results which differed by
several milliseconds.

The material consisted of 8%" x 15" transparent (Cellu-
loid) cardboards introduced in the G backs. These cardboards
supported the fine line to be detected and the dark bars
assumed to exert an influence upon its detection. Thirteen
targets were prepared for the fore period and twelve for the
exposure period. Two more targets, used in exposure period,
were test targets. The first five targets were identical as
for the size of the bars and of the line, the only factor
varied being the distance separating the bars from the line.
In target 1 F (fore period), the two bars, which measured
'i" x'é" (56‘ x 10 12') each, were separated 6" (140 14') on
the horizontal plane, and occupied the middle of the target
on the vertical plane. At the center, two small red disks
of approximately 1/20" in diameter and separated by 5/16"
served as fixation points. Target 1E (exposure) was identi—
cal as for the dark bars, but a fine line made up of copper
wire-é" (1o 12') long and .006" (52") thick occupied the
center of the target and was parallel to the bars. (This
line was cleaned and treated with HgClg in order to become
"gray" and have the least reflection possible, even though
the illumination was coming from behind.) These two targets

are outlined in Figure 1. With this arrangement, the only

29

 

 

 

 

 

 

 

1 F 1.2

Fig. l. Targets 1 F and l E used in fore and eXposure periods respec-
tively. In fore period two dark bars, which measured-fi“ x'é“ (1° 12' x
36'), were 6" (14° 14') apart. Two small "red“ fixation points helped
maintain proper fixation. In eXposure period target 1 F vanished, being
almost instantaneously replaced by target 1 E in such a way that every-
thing seemed unchanged except for the fine line appearing in the middle
of the space previously occupied by the two fixation points. (Not drawn

t0 80310).

50

change observed when one was switching from the fore period
to the exposure period was the vanishing of the red disks
and the appearance of the line in the middle of the Space
separating those two disks. The dark bars for both periods
had to coincide in order to avoid any perceived movement.
Accordingly, great care was exercised during the preparation
of the targets: measurements of the bars and of the length
of the wire were taken by means of a rule graduated in 64th
of an inch, and a template and magnifying lens were used to
facilitate the work. The thickness of the wire was measured
with a Cenco Ratchet Micrometer.

In targets 2 F and 2 E the distance between the bars
was 2" (4° 46') . One inch (20 25'), s" (10 12') and 5/64"
(11') separated the bars in targets 5 (F and E), 4 (F and E),
and 5 (F and E) respectively. In target 5 F the dark bars
occupied the place where the red disks would normally be,
in which case the bars themselves served as fixation points.

In the next group of targets the distance between the
bars was maintained at 5/64" (11'), but this time the length
of the bars was varied. In targets 6 F and 6 E the bars were
4" long (90 51'); in targets 7 F and 7 E they were 2“ long
(40 46'), while in targets 8 F and 8 E they were only i“
long (56'). The length of the line in the exposure targets
(6 E, 7 E, and 8 E) coincided with that of the bars.

In the third group of targets the same lengths were

utilized as in the second group, except that this time the

51

distance between the bars was kept at 6" (140 14'). Thus,
targets 9 F and 9 E contained bars measuring 4" in length
(90 51'), but these were 6" (140 14') one from the other.

Targets 10 F and 10 E were 2" long (40 46'), while targets
11 F and 11 E were-&" long (56').

Target 12 F was a black cardboard with two small fix-
ation points: two small holes were made in the cardboard
and a piece of "red "filter was put behind these fixation
points.

Target 15 F was tranSparent and contained only the two
fixation disks.

Target 14 E had no bars, but only the fine line in the
middle. This line was é“ long.

Two more targets were test targets. One was a piece
of transparent celluloid without any line nor bars on it.
The other was identical to target 5 E, except that it had
no line but only the two bars.

In all the targets the area was kept constant in order
to rule out any explanation of the results in terms of

spatial summation.

Preparation of Subjects

Generalgpreparation. Because the two observers served

 

as experimenters also, it was essential that J.V. as well
as the author be familiar with the psychophysical method
used. As said before, J.V. participated in a first attempt

made in May 1964. At that time the author explained to him

52

the nature and shortcoming of the method to be used and
readings were assigned to him on the topic.

As a subject he was asked to use as consistent a
criterion as possible and he was warned that test targets,
where the line is absent, might be used during the experi-
ment. The author as a subject also was aware that test trials
would be used once in a while when he would be the observer.

No formal written instructions were given the ob-
server, but he was cautioned again and again to look at the
pair of fixation points, and to indicate in the test period
whether or not a line was visible in the direction of the
fixation points.

It is always to be recognized that attention may vary

 

somewhat and that this may be a factor in the production of
variability in the results. In the absence of any marked
variability in the results, it is also possible to account,
in part, for the results as the target is varied, by saying
that attention is different for each target condition.
This is another way of saying that the central nervous
system organization is varied by the targets and that not
all of the systematic differences often attributed solely
to retinal patterns of stimulation and/or to neuroretinal
activity, is rightly so attributed.

Immediategpreparation. At the beginning of each
session the subject dark adapted for 50 minutes; for the

first 15 or 20 minutes he wore dark adaptor goggles (Picker

55

Panoramic Goggle) and then went in the dark room for the

remaining 10 or 15 minutes.

Procedure

The experiment took place in a 16' x 12‘ dark room.
However, two 7.5 watt "red" bulbs were used by the experi-
menter to manipulate the targets, operate the timer, and
collect data. The observer sat comfortably in front of the
apparatus and observed the targets. Before each presentation
a signal was given to him. Between trials, he was informed
not to look directly at the target, especially when the
illumination was high.

Experimental situations. The experiment comprised two
parts. In Part One, which took place in July and August 1964,
six Situations were investigated under three levels of
photic intensity (see Figure 2 for an illustration of these
situations):

(1) The distance between the two dark bars and the line
was decreased. The first five sets of targets were used so
that the distance of the bars respective to one another was
varied in five steps from 140 14' to 11'.

(2) The fore period was dark with only two small fix-
ation points visible, and the line and bars appeared to—
gether in exposure period. Target 12 F as well as targets
1 E and 5 B were utilized.

(5) The fore period was lighted but the target did not

contain the vertical bars. During the exposure period the

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SITUATIONS ‘
FP sp FP EP
1 Eli
32
i 3 I l
13
3 .. Ed 5 I E
13F 53 13F 1E
E! l i H 1
5r 143 1r us
I"H WE Ii IllE
3r 5E 5F 33
III
8F

8 E

 

 

 

Fig. 2. The six eXperimental situations examined under three different
levels of photic intensity. Not all the conditions investigated in the

first and sixth situations are included. I“? is for fore period, E? for
OXposure period. 1 F, l‘E, etc, correspond to the different targets described
previously in the text. (Not drawn to scale).

55

two bars and the line appeared simultaneously. Target 15 F
was used in fore period. The exposure period was identical
with the one in the second situation.

(4) The two vertical bars were present in fore period,
but they vanished and the line appeared alone in exposure
period. Targets 1 F and 5 F as well as target 14 E were
used.

(5) The two bars did not occupy the same position in
fore and exposure periods. In one of the conditions targets
5 F and 5 B were used so that a shifting of borders toward
the center resulted. In the other condition the use of
targets 5 F and 5 E created a shifting of borders toward the
periphery.

(6) The length of the line and of the two bars was in-
creased. Targets 6 F and 6 B through targets 11 F and 11 B
were utilized. The length was varied from 90 51' to 56' in
three steps. Two distances between the bars (140 14', and
11') were used.

Part Two was performed as a check on the reliability
of data in those Situations where the two observers dis-
agreed. The situations re—examined were: Situations 2,

5 and 5. This part took place approximately six months after
the first one, that is in February 1965.

The psychophysical method used was the method of

limits. At first it was decided that the descending as well

as the ascending orders would be used. After a few days of

56

experimentation it was found that the threshold as de-
termined by the descending order was consistently higher
than it was with the ascending order if the illumination was
low and the task difficult. Then the author checked the
literature and found that often the ascending order is used
exclusively. For example, Wilcox used this order only,
arguing that it has "great reliability and consistency"
(Wilcox, 1956). Furthermore, there are other arguments
which can be used against the descending order; for example,
because of habituation one can say he saw the line when in
fact he did not see it, but in the ascending order when he
sees the line for the first time, he is much less influenced
by what he has seen before. Thirdly, the problem of seeing
the after-image is much less critical with the ascending
order.

A minor change in this method was effected: a doubtful
answer was considered as a "no" answer.

Ten threshold determinations were effected in every
case and the threshold of resolution was computed by averag—
ing all the ten readings.

Order of presentation of targets: in Part One, situ-
ations 1, 2, 5, 4, and 5, were tested all together during a
same session at three predetermined levels of illumination.
The purpose, here, was to avoid day-to-day variations
affecting the seeing of the different targets in different

ways. As the order of investigation of each situation might

57

affect the performance of the subject, the order of these
situations 1, 2, 5, 4, and 5, was determined at random,
before each session, by means of a table of random numbers.
Also, for any situation, the order of presentation of each
target size was predetermined in the same manner. For each
target two stimulus threshold determinations was taken in a
row and written down. In some instances where J.V. acted as
subject, three or four determinations were taken in a row.
Also on the first two sessions one or two trials were effected
before recording the threshold values. When all the situations
were covered at the lowest level of illumination, the experi-
menter passed to the next level under a different predeter-i
mined order of presentation. The test targets, without any
line appearing in the exposure period, were used approximately
once every two sessions. A session lasted 5 or 4 hours and
it always took place in the afternoon.

The last situation (Situation 6) was tested alone at a
different session because testing it with all the others
would have required too much time (approximately 5 hours).
In these circumstances one situation had to be discarded.
Situation 6 was chosen because it used targets different from
those utilized in the other Situations. As a result, the data
obtained for that Situation are not directly comparable
with the others.

' In Part Two of the experiment the same procedure was

followed, except that only three situations were examined.

58

As a result a session required about half the time needed
by one in the First Part.

The length of the sessions in Part One is somewhat un-
usual. But it is to be remembered that the subject was in-
structed to ask for a temporary halt any time he felt he
needed one. For this reason the time spent on the task was
sensibly less than the one indicated above. Moreover, a
comparison of the results in Part One and Part Two, this one
being much shorter, does not seem to indicate any detrimental

effect due to fatigue.

RESULTS

For any condition and unless otherwise specified, the
results are the averages of ten threshold determinations of
which no more than two were taken on the same day.

In general the data were stable and consistent.

Table 1 presents in raw form typical data from J.Y.L. and
J.V. on threshold duration for low photic intensity. At
medium and high intensities, the stability of data is even
better.

Table 1. Raw Data on Threshold Duration Obtained in Some of

the SituationS-—for the Two Observers--Photic
Intensity .055 c/ft2.

 

Readings in
Situation Condition Observer milliseconds Means

 

2, Dark fore period Distance: 14014' J.Y.L. 55,65,45,55,45
35,55,55,55,55 so

J.V. 75,95,65,55,55
25,45,45,35,45 54

1, Distance varied Distance: 4046' J.Y.L. 45,45,55,55,55
45,85,85,55,45 53

J.V. 55,55,55,25,15
45,25,55,55,55 54

 

In some conditions J.V.‘s data were less stable and
the task seemed more difficult for him. As this does not

Show in the data chosen at random and presented in Table 1

59

40

above, the two conditions found, upon inspection, to be less

stable than the others are reported in Table 2.

Table 2. Raw Data Found to Be Less Stable, Photic Intensity
.055 c/ft2. Observer J.V.

 

Readings in
Situation Condition milliseconds Means

 

2, Dark fore period Distance: 11' 205,155,565,265,95
65,245,505,175,85 196

5, Shifting of Distancg in 285,175,245,225,85
borders fore: 2 25' 85,145,125,85,65 152

 

First Situation. Effect on visual acuity of a decrease
in the distance between borders.--Figure 5 illustrates the
effect of a decrease in the distance between the two dark
bars upon detection of the line. The results of J.Y.L. are
shown in Figure 5A. First, it can be seen that for any dis-
tance examined, the time required for seeing the line de-
creased as the illumination was increased from .055 c/ft.2
to 10.97 c/ft.2. Second, the effect of the distance between
the bars upon the threshold of detectability is clearly
apparent at all the three levels of intensity used. A Kendall
coefficient of concordance (W= .85) is significant at P < .02,
which means that for this observer the target (distance) that
required more time under low illumination also tended to

require more time at medium and high photic intensity.

Mindanaivwvfiwm“ H._.H F...-

7— H

wt

A~.H~dm_u\n1~v-~ur

HA ed H . H. On.-.~.u.~A H

as dwuva.“

 

MEAN DETECTION THRESHOIDS IN MILLISECOIDS

41

 

 

 

80,,
, l B
i DISTANCE BETWEEN BARS
70-1 4‘ \ '— _ "‘ “ 11'
i \ e— — -- -I 1012'
1 : ‘ .____g 2023'
g , \ Ar— — — at 4°46.
.0; I \ ~———- 14%
i
l
3
50-3
i
40a
304
20.

 

 

J.Y.L. J.V.

 

 

 

.oéa .19 15197 ' .053 .33) 173.97

marques (c/rr.2)

Fig. 3. Visual acuity as dependent upon the distance of adjacent
borders. The distance between the inner borders of the two vertical
bars was varied in five steps from 14° 14' of visual angle to 11'.
In part A, JeYeLe V88 the observer, and JeVe in part.B.

42

However, the duration required for detectability of
the line did not necessarily increase proportionately with
a decrease in distance from 140 14‘ to 11'. If one takes
the curve obtained for the condition where the distance
between the bars subtended 140 14' as the criterion for
comparing all the other curves it can be seen that the fine
line was perceived more rapidly at 10 12' and 20 25' than at
140 14', the condition where the bars were almost coincident
with the frame and as such were assumed not to exert any
influence on the detectability of the line. This effect is
maintained at all three levels of photic intensity, thus
Showing some facilitatory influence of the bars upon visual
acuity.

When the bars were very close together (11'), however,
it took much more time to detect the line, as can be seen
by inspecting the upper curve of Figure 5A, which indicates
some inhibitory effects of the bars upon visibility of the
line. Also, it can be seen that while the effects noticed
for all the other distances examined tended to diminish
when the photic intensity was increased to 10.97 c/ft.2,
the inhibitory effect created by the bars when they were
close to one another tended to be comparatively larger at
the same intensity.

Results of J.V. are shown in Figure 5B. It can be seen
that data for this observer are very similar in trend to
those of J.Y.L. A Kendall W test (w= .79) was significant

at p < .05.

45

As compared to the standard condition, the inhibitory
influence of the bars when they were close together is even
more apparent here. Yet, the facilitatory influence noticed
above is much less evident in Figure 53. For example, at
10 12‘ duration is increased as compared to the standard
condition, thus indicating that some inhibitory influence
began to appear when the bars were slightly more than 10
apart. At a greater distance (20 25') and with the lowest
illumination used there is a slight indication of the
facilitatory effect noticed before.

The curve corresponding to a distance of 40 46' is
almost identical with the criterion, thus showing some sup—
port to the assumption that at a distance of 40 the influence
of borders upon one another is no longer effective.

For this observer also, while the effects of borders
at 10 12', 2O 25', and 40 46' seemed to be less marked when
the illumination was increased (the curves are very close
together at 10.97 c/ft.2), the inhibitory effect of borders
at 11' was maintained even at the highest illumination used.

Situation 2. Effects of a darkened fore period.--
Besides distances between borders, other manipulations were
effected. For example, in one of the Situations the fore
period was dark, with only two small fixation points visible
and the line and bars appeared together in exposure period.

Figure 4A presents the results obtained by J.Y.L.

while those of J.V. are shown in Figure 5A.

MEAN DETECTION THRESHOLDS IN ldILLISECOICDS

44

 

 

 

 

 

 

 

 

 

 

 

60 a 11.0..1
r A _, c
1.0 - 120 j
DISTANCE BETWEEN BARS
\
1 -( \ e-— -— ——e 11'
20 . 1004 \ 11.011.‘
.4 -
J.Y.L.
0 V 1 fi 80 -1
so -,
40-1
20 1
J.Y.L. J.Y.L.
o . , 11 o , fi .
.053 .69 10.97 .053 .65 . 10.97

mmnscs (c/r'r.2)

Fig. 4. Effects of diverse manipulations of adjacent borders upon
detection of the line. Data for J.Y.L. In part A the fore period was
unilluminated; in part B it was lighted but contained no vertical bars,
while in part 0 the bars vanished in eXposure period.

In.

dhuA ne— A w\.‘v4.—... ..

1‘

—|..-.. N 1nd

w.—

H

aha—KNAVu—

Q.“

.nuv—u THU

whAU

H.

rv‘Vu.~.Huu.HnL—.

z < Inn .N

46

By examining Figure 4A one can see that at low illumi—
nation it took less time for the observer to detect the line
when the bars were close together (11') than when they were
far apart (140 14'), suggesting that in some conditions
borders near one another exert a facilitatory influence upon
visual acuity. At medium intensity no difference was ob—
served between the two curves, thus indicating that the
distance between the bars did not interfere with visual acuity,
while at high intensity there seemed to be a tendency for
both curves to raise again, this tendency being slight for
the curve representing a distance of 11' but more accentuated
in the case of the wide distance.

Data for J.V. show a very different pattern depending
on the distance of bars (Figure 5A). Thus, if at medium
intensity the time required for detection of the line was
about the same be the bars close to one another or far apart,
one cannot conclude the same thing when low or high illumi—
nation were used. At low intensity much more time was
required for detecting the line when the bars were near one
another. This is the opposite of what was found with J.Y.L.
as the observer. At high intensity the trend is the same
as the one observed in J.Y.L.'S data, except that the tendency
is much more pronounced here; thus, it became much harder
for J.V. to detect the line when the bars were far apart.

It seems as though in the present situation, borders in near
the vicinity of the line improved visual acuity by sub-

stantially decreasing the glow affect produced under some

47

conditions when the bars are far apart. This will be further
discussed in the prOper section.

Because the data of the two observers did not agree,
another session took place six months later (February, 1965)
in order to check their reliability. In Figure 6A it can be
seen that the first observer was fairly consistent over a
period of six months. The same tendency is observed. Only
for the lowest intensity is there any increase in duration
noticed, Specially in the condition where the bars were close
together.

Equally, the second observer (J.V.) was consistent over
the same period. By inSpecting Figure 6C one can notice that
the trend remains the same, except that the difference be-
tween the two curves representing wide and Short distances
respectively were less marked at low intensity than on the
first occasion.

Situation 5. Effects of a lighted fore period without

any borders.--In the present Situation the fore period was

 

lighted but the target did not contain the vertical bars.
During the exposure period the two bars and the line appeared
Simultaneously. In one condition, the bars were 140 14'
apart, and in the other they were close together (11'), as

in the previous situation. The curves obtained can be seen
in Figures 4B for J.Y.L. and 5B for J.V. In Figure 4B one
can again notice the tendency for visual acuity to require

less time with an increase in illumination. However, it does

g‘

«.4
nnfldéflvuwfhun H. u..~. H414

wk

H

Hung ~.~.fiJ——MU«V~U——u.fih ZAJH.<.HN.~.H.JHAN

7N < 4. .mm‘

MEAN DETEETION THRESHOLDS IN MILLISECONDS

 

 

 

 

 

 

 

 

  

 

 

 

100 1
D
80 T
60 . DISTAhOE BETWEEN BARS
.. } 140 14!
1.0 1 o ___________ o} 11
J
20 - \\
- \
J.Y.L. \
0 . Q \
I“‘ \
.“ \
80 l '. \
\
-< X. \
60 4 \
J \
.“ \
4o 4 ‘ \
20- .0
J.Y.L. J.V. J.V.
0 I T I O r j —1
0053 .69 10.97 .053 .053 .69 10.97
LUMINANCE (c/rr.2)
Fig. 6. Comparisons of sets of data collected at a six month interval. Portions

A and C represent, for J.Y.L. and J.V. respectively, the situation where the
fore period was dark, and portions B and D stand for the one where the fore pe-
riod was illuminated but contained no bars. The solid lines represent the con-
dition where the bars were far apart, and the broken ones where they were close
together. The’closed circles represent data gathered in July 1964, and the open
circles data for the same conditions collected six months later.

49

not seem that the presence or absence of borders near the
line differently affected visual acuity when these borders
were presented at the same time as the line. A Sign Test for
paired replicates was not significant at p= .05.

If one looks at Figure 5B, he can observe a clear dif-
ference between the curve representing data when bars were
close together and the one for a distance of 14C 14': thus,
it took J.V. much more time to see the line when the bars
were close to it than when far apart. This is not in agree-
ment with results for J.Y.L.

Because the two observers did not agree when tested on
the conditions depicted here, a retest was effected six
months later. The results obtained on the second occasion
appear in Figures 6B and 6D. Both observers were consistent
over that period, the only difference worth noting being
for J.V. a certain improvement in detecting the line when
the bars were close to it, and for J.YLL. a slight but con-
sistent increase in time for the three levels of photic
intensity, as one can see by comparing the curves drawn with
open circles with those drawn with closed circles.

Situation 4. Effects of disappearipg_borders upon

 

detection of the line.-_Another situation was one in which

the vertical bars, which were present in fore period, vanished
and the line appeared alone in exposure period. The results
for J.Y.L. are shown in Figure 4C, and those for J.V. in

Figure 5C. Here the two observers Show a very similar pattern

50

of data. For the two of them there was an improvement in
visual acuity when the intensity was increased from .055
c/ft.2 to 10.97 c/ft.2. For the two observers also,
detection of the line required more time when the borders
of the bars were near one another, thus suggesting a strong
inhibitory influence in such a condition.

Situation 5. Effects of a shift in the position of

the bars in fore and exposure periods.--In the present Situ—

 

ation, the dark bars occupied a predetermined position in
fore period and a different one in exposure period. As the
switching from fore to exposure was almost instantaneous,
there was a shifting of borders from one position to another.
In one of the two conditions examined the bars were set at

20 25' from one another in fore period and at 11' in ex-
posure period. In the other, the reverse was done. The
results obtained with these manipulations are shown in
Figure 7. In portion A the curve represented by closed
triangles has been obtained when the borders were 20 25'
apart in pre—exposure and 11' in exposure periods. The
closed circles represent the curve obtained when the borders
were 11' apart in pre—exposure and 20 25' in exposure.

A comparison of those two curves indicates that when borders
were Shifting away from the line, this one required much
more time to be detected than when they were shifting
toward it. A replication of Situation 5 Six months later

gave the two curves represented by open circles and open

mu.~.~.ew.w.:nh.—.~HT- 2H qu~.~A::n.&:—:.F .nn,.r.rr..h....n;> 2...;i.‘

 

 

 

MEAN DETECTION THRESHOLDS IN MILLISECONDS

160-1L
9
140 d .e‘
1204 \
‘
100— \
\

 

201

i

.053

Fige 70

l
.69

51

  

j
10.97

 

—l

—

(DISTANCE bsrwzsn
BARS IN EXP.FERIOD)

 

    

1...“.

 

_ '_""”"" } 2° 23!
o- ---------- -e
J.V.
.653 .65 10397

LUMINANCE (C/FT.2)

Visual acuity as affected by a shifting of position of adja-

cent borders in fore and exposure periods. Portion A represents data
for JeYeLe and portion B data for JeVe
shifting of bars toward the line, while the broken lines represent

a shifting away from the line. The Open triangles and circles are for
the retest, made six months later, of the conditions represented by
closed triangles and circles. The crosses display data showing poor
stability (see text for eXplanation).

The solid lines represent a

52

triangles. Again the stability of the results is well main—
tained, the tendency is the same, the only difference being
a higher threshold for the two conditions where photic
intensity was .055 c/ft.2. For medium and high intensities
there is an overlapping of curves obtained over the period
of six months.

Data for the second observer are illustrated in part B
of Figure 7. The results yielded by the condition where
the shifting was toward the line (identified by crosses)
and the condition where it was away from it (identified by
closed circles) were about the same- At that time (July 1964)
an inspection of the data indicated a noticeable tendency
toward a decrease in duration as the experiment went on
(Table 2, p. 40). To check that point the A. took several
more data for this situation as well as for some others.
While the tendency remained the same for the other situations,
it definitively changed in the condition where shifting was
toward the line. The curve identified by closed triangles
shows other results based on eight threshold determinations
for low and medium photic intensities, and six determinations
for the high level of intensity. A valid explanation of
this marked decrease in duration is not easy: the targets
have been carefully inSpected and there was nothing wrong
with them. One possibility is that the observer did not
fixate properly at first; he may have been distracted by the
apparent movement of the two bars and may have learned,

after several trials, to look only at the fixation points,

55

not the bars. To check that assumption, J.Y.L. tried, later
on, to bear his attention on the bars and he found out that

apparent movement became more objectionable and that visual

acuity was much impoverished.

A replication of the whole situation conducted six
months later showed essentially the same trend for the con-
dition where the bars were set at 20 25' in exposure period
(curves with closed and open circles). When the bars were
close together in exposure period the data were almost
identical with those obtained as a check and based on eight
determinations (curves represented by open and closed tri-
angles in part B of Figure 7). With this taken into consider—
ation it can be seen that the two observers agreed in their
tendency to require more time for seeing the line when the
borders of the bars shifted away from it.

Finally, an overall inspection of Figure 7 shows that
whatever process was acting at the time, there was a marked
decrease in duration required for detection of the line as
photic intensity was increased.

Situation 6. The effect of length of the line and
bars upon visual acuity.--In the present situation three
different lengths of bars and line were used. Portions A
and B of Figure 8 present the results for the first observer
(J.Y.L.) and portions C and D present those for J.V.

In part A the distance was kept at 140 14'. The length

of the line here became a significant factor in visual acuity,

MEAN DETECTIOI‘J THRESHOLDS IN MILLISECOIDS

54

    

 

 

LENIfl1CW'BARS
60~ ‘ . A . x, B x—————at9°31'
40 - “.\ ..
- K ‘ _
\

20 d ..

Distance 14° 14' l Distance 11'

J.Y.L. J.Y.L.
O I l I l I I 1

 

 

60 ‘1 -
x .1
‘\
40 -4 “x ..
\
«r "x \ ‘ ‘ -
\

    

‘ Distance 140 14' “ Distance 11'
J.V. J.V.

.653 .5 16197 .353 .89 £9.97

 

 

 

 

MIINANCE (c/FT.2)

Fig. 8. The relation between bars length and visual acuity for two
different distances of the here. In parts A and G the two vertical
bars are kept at a distance of 11.0 14', while in portions B and D
they are at 11' of visual angle. Data for J .Y .L. are displayed in
parts A and B and those of J .V. in parts 0 and D.

55

eSpecially at low intensity where the required duration
decreased With an increase in length from 56' to 90 31'.
This is in agreement with results of Hecht, Ross & Mueller
(1947), and Ogilvie & Taylor (1959). It can be seen that
when the length does not subtend at least 10 the threshold
of detection raises sharply with a decrease in illumination,
as Hecht and collaborators have found.

In part B, the same bars and line were used, but the
distance between the bars was kept at 11‘. A comparison of
the three lengths used does not show any differential effect
due to the length of the line and bars. It seems as though
there was a slight reduction in required duration when the
line and bars were very short, but this is about all. The
six conditions (three lengths at two different distances)
have been submitted to the Kendall coefficient of concord—
ance (W) and this one has been found to be significant at
p < .05. So, the rank of each condition tended to remain
the same over the three levels of photic intensity used.

Figure 9A, B, C, is a different illustration of the
phenomenon. Here, instead of comparing the effects of a
difference in length upon detection of the line, the A.
compared the effects of distance. Hence, part A of Figure 9
represents the influence of borders of long bars (90 51')
upon visual acuity first when these bars were near one
another (11'), and second, when they were far apart (140 14‘).

Part B illustrates the same phenomenon for bars of medium

ECTION THRESHOLDS IN MILLISECOIJB

MEAN DET

56

  
  

  

 

 

 

 

 

       

 

 

 

65., q
60‘ DISTANCE 51;?me ems ’ "~ \
‘ \.
x\ F--—-——x ll' \\\
.4 \\ F—‘X 140 14c .. \
\\ X\
40- \ q \
\\ A \ D
. .K‘ \\
\~ “ \
20~ ‘x e
d) J'YSI" ‘ J.V.
Length 9° 31' Length 90 31:
o . , fl . . fi
60- _
x l
\
- \\ B d E
\
\ Xf— _____ —- x
40- \\ - \\\
q ‘\\x\ ‘ ‘\
“~x
20' -
J.Y.L. ‘ J.V.
Length 4° 46' Length 4° 46'
o _

l A
' I ‘1 v v ~—j

 

   

 

 

 

 

60-, .
F
40‘ -
20~ a
‘ ‘ J.V.
Length 36' Length 36'
0 j l I ~——l

.053 .89 10.97 .053 .69 10.97

matures ( C/FT.2 )

tance between
F1 . 9 Visual acuity as dependent upon bar length and dis

ths bars. In parts A and D the length of the bars and line subtends 9° 31',
in parts B and E 4° 46', and in portions C and F 36'. In each case the
broken line represents the narrowly spaced bars (11'), the solid line,

the widely spaced ones (14° 14').

57

length (40 46'), and portion C for short bars (36'). An
overall inspection of the three graphs shows that when the
bars were close to the line, whatever the length of the line,
this one required a fairly long time in order to be detected.
When the bars were far apart, however, the duration required
decreased as the line extended. Finally, the facilitating
effect of an increase in luminance is clearly apparent in
this situation also.

Data obtained when J.V. was the observer are in
general similar to those of the former observer. A Kendall
coefficient of concordance is significant at p < .01.

Figure 8C shows the effects of increasing the length of the
line when the bars were very far apart (and considered to
be nonexistent). At low illumination eSpecially the time
was reduced when the length was increased.

Figure 8D resembles Figure 8B for medium and high
intensities. At low illumination, however, duration was
increased for seeing the long line as compared to the other
two lengths.

Parts D, E, F show the same trend notices for J.Y.L.,
that is when the bars were far apart, the length of the
line seemed to be a decisive factor for its visibility.
When the bars were close to it, time was comparatively in-

creased. While for J.Y.L. this effect was manifested at all

three levels of photic intensity, it was more pronounced at

medium intensity in the case of the second observer.

58

To this point, the results yielded by the six experi-
mental situations have been reviewed separately. Figures
10 and 11 present an overall comparison of all the situations,
with the exception of Situation 6 which, using different
lengths of line and bars is not directly comparable. Thus
in Figure 10, which represents data for J.Y.L., a comparison
of Situations 1 and 2, when the bars wereofar apart (first
and second blocks of columns from the left) indicates that
a darkened fore period shortened the time required for
visual acuity at medium intensity only: at this level it
took the observer 5 milliseconds to detect the line when the
fore period was unilluminated as compared to 22 milliseconds
in the reference situation (Situation 1).

When the distance between the bars subtended ll' (fifth
and sixth blocks from the left), it became much easier to
detect the line when the pre-exposure was darkened than
when it was illuminated and contained the two bars. This
facilitatory effect was maintained over the three levels of
intensity used.

There was not much difference, except at low illumi-
nation, between Situations 1 and 5 when the distance between
the bars was wide (first and third blocks from the left),
but when it subtended only 11' (fifth and seventh blocks)

a lighted fore period without any borders required less time
for detection of the line as compared to a fore period with

the presence of borders. These results suggest that borders

59

.A.w.h you open .pOHsoQ ohdmOQXo ca pom: one one ma poosoacca conspmfip one
.m soapsSpflm mom .mcoapsspwm anacoaauuexo say we o>am you mvHonmousp cowpsnso mo comwpsdsoo .OH .mah

m z o H e d D a H m

w H e m

 

as m

SGNOOESITTIW NI SGHOHSHHHL NOILOHESH NVSW

 

 

 

 

mb
ZDHQHE WWW
Om. .
mon flHQ
“OH
whHmszZH
r A‘l \ r 1W \ r 1‘ L ONH
MN 0N .HH .VH OQH
«ma

mm<m zmmjemm Moz¢BmHn

60

in near vicinity but appearing together do not exert the
depressing effect noticed when some of these borders appear
before others (bars presented before the line).

In Situation 4, where the bars vanished in exposure
period, the block representing the condition when the bars
were far apart (fourth block) resembles the block for the
reference condition (first block). This is an expected
result because in the present situation as well as in the
standard one there were no borders in the vicinity of the
line. When the bars were close together, Situation 4 re-
quired more time than Situation 1 (fifth and eighth blocks).
This indicates that borders in close vicinity which vanish
while new contours are forming exert a greater inhibitory
influence on these than borders which are stable.

In Situation 5, if the condition where the distance
between the bars was 11' in exposure period (ninth block)
is compared to the reference situation (fifth block from
the left), one can observe an improvement in visual acuity
with a shifting of the bars toward the line. This improve-
ment is maintained over the three levels of intensity used.
When the distance, in exposure period, subtended 2o 25',
shifting of the bars away from the line greatly impoverished
visual acuity, as compared to using steady borders (the
two last blocks to the right).

An overall inspection of this figure shows that thres—

hold duration for detection of the line was comparatively

 

 

61

longer when the two bars, near one another in fore period,
disappeared simultaneously with the appearance of the line
(Situation 4), and also when the bars, close to the line
in fore period, were shifting away from it in exposure
period (Situation 5).

Figure 11 presents a comparison of data for the second
observer (J.V.). For example, a comparison of Situations
1 and 2 (first and second blocks of columns) indicates that
when the bars were far apart, a darkened fore period com—
paratively impoverished visual acuity in all three levels of
intensity, but the effect was much more noticeable at high
intensity. This is contrary to what has been found when
J.Y.L. was the observer. When the bars were close together
it can be seen that at low illumination a dark fore period
required much more time comparatively (fifth and sixth blocks
from the left). The reverse was true at medium illumi-
nation, while no difference seemed to exist at high illumi-
nation.

When a lighted fore period without any borders was
used (Situation 3), and bars appeared at 140 14' one from
the other in exposure period, the results were almost
identical to those yielded by the reference condition
(first and third blocks). So, for this observer as well as
for the previous one, when the bars were presented with the
line but at a great distance from it, the result did not

differ from what was obtained when the bars were present

 

Ulainuub DDLWL‘AL'JJV umu

62

 

3

R\\\\\\\\\\\\\\\\\\\\\\H\\\\\\\\\\\W

2°2

  

 

mm _}
me \\\\\{Q\U{\U\\WQ

III IIWIIIIIIIIII
m\\\\\\\\\\\\i\\\m

@¥E~
EW\\Q\\\\\\

I §\\\X \H

 

 

 

 

14c 14-
INTZNSITY
D HIGH
g m
s w
a? 4

yeah detection thresholds for five of the situations, with J.V. as the observer. For

 

 

I l l l l T l j
0 tn 0 o n O In 0 In 0
0 u‘ on «v E; 0‘ e~ .9 an F1
N H H H H \t

SGHOOHSITTIH NI SGEOHSHHHL NOILOELSG NVSH

.
I IIIIIII III IIII IIIIIIIII IIIHIIIH IIIIIII IWI]-

S I T U A T I 0 N S

the distance indicated is the one used in exposure period.

5:

gig, ll.
Situation

65

in fore period. This suggests that at such a distance the
presence or absence of borders does not exert any influence
upon visual acuity.

This observer, however, contrary to J.Y.L., required
more time to detect the line when the two bars, absent in
fore period, appeared with the line and close to it in
exposure period (fifth and seventh blocks).

In Situation 4, when the bars were far apart the re-
sults were very similar to those yielded by Situation 1
(first and fourth blocks). This is what was obtained when
J.Y.L. was the observer. But when the bars were close to—
gether, the present situation in which borders vanished in
exposure period, required much more time than Situation 1
(fifth and eighth blocks). Again this is in agreement with
results found for the other observer.

A comparison of Situation 5 with Situation 1 clearly
indicates that at low and medium intensities, a shifting
of the bars toward the line required more time for its
detection than Situation 1, where the bars remained close
to the line (fifth and ninth blocks). This is contrary
to what was observed in the case of J.Y.L. A possible ex—
planation is that if, for any reason the line can be detected
before any apparent movement is seen, then shifting of the
borders toward it improves visual acuity. But if it takes
so much time to detect it that some apparent movement is

first perceived, this phenomenon will make detection of the

 

64

line more difficult. This will be explained more fully in
the appropriate section.

When the two bars were shifting away, being set 20 25'
apart in exposure period, one can notice, as was the case
with J.Y.L., that visual acuity was greatly impoverished in
this condition, compared to what was obtained with the same
distance in Situation 1 (the last two blocks).

A comparison of Figure 11 with Figure 10 indicates
that for this observer as well as for J.Y.L. visual acuity
was markedly impoverished in Situation 4 where borders near
one another disappeared simultaneously with the presentation
of the line, and in Situation 5, when the bars, close to the
line in fore period, were shifting away from it in exposure
period. However, in the case of J.V., Situation 2 com-
paratively required much more time than the same situation
for J.Y.L. Thus, when the bars were close together (sixth
block) a low illumination greatly impoverished visual acuity
in the case of J.V. Equally, when the bars were far apart

(second block) a high illumination yielded the same result.

 

 

DISCUSSION

The purpose of the experiment was to examine diverse
spatial and temporal manipulations of target borders upon
detection of a fine line. The dependent variable was dura-
tion of stimulation. The main assumption, underlying the
whole research was that, as contour formation takes time
to achieve completion (Werner, 1955), variations in duration
would be an excellent index of facilitatory or inhibitory
effects of contours upon one another. As compared to some
standard condition, an increase in duration would indicate
depressing or inhibitory effects, while a decrease in dura-
tion would indicate a facilitatory effect. In a certain way
the results of the present study are not directly comparable
with those found when time was unlimited or sufficiently
long so that it did not become a significant factor in the
outcome. As duration is a very important variable in any
neurological process, it was thought that it would allow
one to better understand the processes at work. For example,
with a target presented for approximately .2 second (which
is relatively brief anyway), no difference would be found
between any of the variables investigated in the present
research. However, when time becomes the dependent variable,

the results have shown that inhibitory, and occasionally

65

66

facilitatory effects could be clearly demonstrated. For
this reason it is the belief of the author A. that time
should be more extensively utilized as a dependent variable
in studies on visual acuity than it has been till now.

Distance of borders.--Studies on visual acuity have

 

shown that borders interacted. Yet, this interaction was
found to be limited to a relatively small distance, that is
around 4' (Fry & Cobb, 1935; Kravkov, 1958). Targets usually
consisted of a single bar or line of varying width, or of
twin parallel bars the interSpace between which was to be
detected.

In brightness discrimination, the interaction was
found to cover relatively large areas (Fry & Bartley, 1955).
The effect was found to be inhibitory when borders were
parallel to one another, and facilitatory when they acted
at right angle upon one another.

The present study shows that when a line is to be
detected there is an inhibitory effect of borders at a com—
paratively short distance, and at least for one observer
a facilitatory effect at a greater distance.

In order to see the importance of the inhibitory
effect of borders in near vicinity, the length of the two
vertical bars and of the line was increased in Situation 6.
It has been found in the past (Hecht, Ross & Mueller, 1947;
Ogilvie & Taylor, 1959) and in the present research when

the bars were far apart, that when it is presented alone a

 

 

 

 

 

67

longer line is more easily perceived than a shorter one.
So, the length of a line has a facilitatory influence upon
its detection. On the other hand, borders close to it have
an inhibitory effect. The results indicate that the
depressive effect of borders was progressively increased
and it compensated for the facilitatory effect of length.
For one observer, approximately the same duration was re-
quired to detect any of the lines when the vertical bars
were close to it. For the other, the longer line required

even more time indicating that the inhibitory influence of

 

borders was increasing more rapidly than the facilitatory
effect of length.

The principle stated by Fry and Bartley and already
utilized elsehwere (Fry & Cobb, 1935; Kravkov, 1958) seems
applicable to explain this depressive effect of borders at
short distances. Therefore, the processes into play are
assumed to function in the following way. The two parallel
bars lead to the formation of their respective contours
during the fore period. In exposure period, when the line
appears, the contours corresponding to its projection on
the retina are depressed by those contours already present,
and more time is required for these new contours to complete
and the line to be seen.

More Specifically, when the bars are near one another,
their action can be analyzed in this way: first, being

much wider than 4', their reSpective inner and outer borders

 

68

do not interact significantly. If the retinal image of

these bars were a perfect replicate of their physical proper-
ties, their projection would cover approximately 180 microns
in width, one minute corresponding, in the typical emmetropic
eye, to a retinal distance of approximately 5 microns (West-
heimer, 1963). But because their projection is Gaussian

(Fry & Cobb, 1935) or approximately so (Westheimer, 1963), it
will spread over many more microns and there will be an over-
lapping of the two gradients. From this point onward, and
for the very reason that neurological studies on contour
formation are still in their infancy, the author cannot but
provide a very general description of what he believes to
take place. It is assumed that detection of the line and

the bars depends on the activity of off elements in the
retina or elsewhere along the pathway. With a sudden decrease
in illumination, as is the case when the bars are projected,
off cells are activated. However, not all the off cells
reached by the gradient—wise projection of the bars are
affected in the same way. Below the point correSponding ap—
proximately to the border of the bar in the gradient (Figure
12), there will be an increased activity of off elements,
while above that point these off elements will be inhibited.
Thus the difference in activity on either side of the border
is augmented and leads to the perception of a sharp edge.
Some evidence for such a process at the cortical level is

presented by Jung (1961). When the line is presented, its

 

 

69

 

 

 

Fig. 12. Gradientawise projection of a dark bar upon

the retina. Solid lines represent distribution of radia-
tion in the target; broken lines, the retinal distribution
of radiation. The part of the curve below the two points
corresponding approximately to the borders of the bars ( x
and x') is characterized by an increased activity of off ele-.
ments, while above these points there is an inhibition of
activity of the same type of receptors. Thus the perception
of a sharp edge depends on the difference in activity of off
receptors above and below the transition points. (Note: the
transition points are not necessarily exactly at the juncture
between the solid and broken lines. There are conditions ,
(Kravkov, 1938) wherein these points are shifted on a lower
part of the inversed curve and, as a result, the bar is per-
ceived as thinner).

7O

contour processes have to form in that area of overlapping
gradients where inhibition of activity is relatively great.
Therefore, it will take time for the contour of the line to
overcome that inhibitory influence and reach completion.

When illumination is increased, the line contours are
affected in three ways. First, the contours are enhanced
because, the gradient becoming steeper, the difference
between the two types of activity of off elements at the
borders becomes greater. But, the line being narrow, its
two adjacent longitudinal borders are near one another and
they interfere more and more in the building up of their
reSpective contours as intensity is increased. Perceptually
the line is seen as thinner. The facilitatory effect of an
increased steepness of the gradient, however, overcomes the
depressive effect of short distance between the line con—
tours and detection is facilitated (Kravkov, 1938). Third,
the bars also have much steeper gradients, the activity of
off fibers, below the transition point is much greater and
inhibition should be more serious above the transition point.
As a result the facilitatory effect of increasing illumi-
nation upon detection of the line would be more or less
counterbalanced by the contours having to build up in an
area of greater inhibition.

When the distance between the retinal projection of
the inner borders of the two bars covers about 10 12' or

360 microns, the results obtained (at least for one observer)

 

71

imply that the line contours are facilitated. Thus the
mechanism is different from what has been proposed to ex-
plain inhibition. Nothing in the studies on the relation
between contour formation and visual acuity can be helpful
in this particular case, because the distances investigated
were much smaller, subtending a few minutes of arc. There-
fore suggestions concerning the mechanism at play have to be
sought elsewhere. The structure of the retina, for example,
may be thought to have something to do with the outcome.
According to fisterberg (1935) the distribution of cones in
the fovea decreases sharply as one goes from the center to
about 30 peripherally. Thus, at a distance of 10 from the
center there are more than twice the number of cones per
unit-area than at a distance of 20. The A., however, does
not see how the structure of the retina, per se, would
create facilitatory conditions for the formation of contours.
If so, the facilitatory effects should be even greater when
the bars are very close to the line, which was found not

to be the case.

Another possibility would be that the area of depressed
activity mentioned above is succeeded by one of enhanced
activity of off fibers, and the line contours would build up
at the location where facilitatory effects are maximal.

Thus line contours would reach completion in less time and
the line would be seen faster. This is highly conjectural
and the author does not know of any neurological evidence

to support it.

 

72

The facilitatory effect discussed till now was much
less evident in the case of J.V. As it was discovered after
the experiment proper that this observer was suffering from
a mild degree of hypermetropia, it is not impossible that
this defect of refraction, by changing the shape of the blur
gradient concealed the other phenomenon.

To summarize the discussion to this point it can be
said that even in the fovea there is a long range interaction
between processes developing Spatially, and this type of
interaction does not seem to depend on the same mechanism as
the short range interaction. Further psychophysical
researches, as for example variations in the size and location
of bars, in their length respective to the length of the line,
etc., and also neurophysiological researches are needed before
any precise mechanism can be described.

Absence of borders in fore period.-—Another question
worth of interest was this one: If previously formed contours
exert a depressing influence upon the building up of new
contours in their immediate surround, what would happen when
all borders are presented at the same time? It was expected
(Situation 3) that the bars contours, forming at the same
time as the line contours, would depress these to a lesser
degree than when previously formed. More specifically, the
bars and the line contours, having to build up simultaneously
would interfere. The bar, however, being wide, would normal—

ly lead to stronger contours, and the line, on account of its

 

 

73

being narrow, would give rise to comparatively weak con-
tours, so that any interference would be to the detriment of
the line contours. But as the bars contours are forming
simultaneously with the line contours, they would inhibit
these less than if they had achieved completion before the
line contours begin building up.

That expectation, however, receives only partial sup-
port. In the case of one observer, less time was required
for detection of the line when the bars were presented Simul—
taneously with the line and close to it than when they were
projected in fore period. But, contrary to this there was a
depressive effect for the other observer. Those two kinds
of data are difficult to integrate in a simple explanation.
Further research is needed before one can elucidate the
conditions wherein eachane of these effects is at play.

When the fore period was dark (Situation 2) again no
borders were present. This situation, however, is very
different from the previous one and it may lend itself to
complex phenomena an explanation in terms of contour for-
mation will have to deal with.

For one, illumination is much more effective in such
a situation. To be Specific, the retinal elements activated
by a steady illumination do not fire maximally and repeatedly,
they alternate in their activity, as Bartley has explained
(1958b, 1961; Bartley & Nelson, 1963).

But, when brief individual pulses (33 and 66 milli-

seconds) are used, it has been found by Bartley and

 

 

74

collaborators (Nelson, Bartley, and Jewell, 1963) that these
pulses, when separated by null periods of slightly more than
.two seconds, were seen as brighter than a steady illumi—
nation of the same intensity.

The present situation in which the fore period was
dark and the whole target, with bars far apart, was suddenly
highly illuminated for a brief period of time is comparable
to the Situation investigated by Bartley and co-workers.

For example, the duration of presentation of the target, in
J.Y.L.'s case, lasted on the average 13 milliseconds before
he could detect the line, while it lasted 154 milliseconds
in the case of the other observer. One can deduce that if
for any reason the line is not detected almost immediately,
that is when intensity iS still fairly inefficient (Bunsene
Roscoe Law), a duration will be reached which leads to
brightness enhancement, due to all or almost all of the
available elements firing together. This will interfere
with transverse or lateral activity necessary for visual
acuity (Bartley, Nelson, and Soules, 1963). The line will
be seen only when duration is long enough for the retinal
elements not to fire maximally and all together, but to
alternate, permitting resumption of activity across the dif-
ferent channels. Thus brightness enhancement will be
responsible for the wide gap between 13 and 154 milliseconds.
It is worth mentioning, at this point, that J.V. often com—

plained that he was dazzled by the presentation of the

75

target and asked the Experimenter to wait longer between
successive presentations.

When the two bars were close to the line, however, a
well marked decrease in duration was observed at high in-
tensity in the case of J.V. and a slight one in the case of
J.Y.L. (for this observer the threshold was already very low
and a significant decrease could not be expected).

These results imply that the lateral activity of strong
contours (bars contours) in the vicinity of the line contours
present longitudinal activity in the different channels from
being maximal. Thus, if brightness enhancement can depress
weak contours (as is the case in visual acuity), strong
contours in proper location can inhibit brightness enhance-
ment. This hypothesis is certainly worth being investigated
more thoroughly.

Puzzling results are the ones obtained by the two ob—
servers at low photic intensity. Here again, visual acuity
was facilitated for J.Y.L. while it was greatly inhibited
for J.V. As illumination was low,‘brightness enhancement
cannot be reSponSible for the longer duration required by
J.V. Brief, when there are no borders present in fore period,
be this one illuminated or dark, the outcome is this: one
of the observers requires less time in both situations than
when borders are present in fore period, while the other
requires more time. In presence of such contrasting results
the A. can only recommend that more researches be performed

on the problem.

 

76

Absence of bars borders in exposure period.--The re-
sults clearly suggest that contours which come undone while
new ones are forming up in close vicinity exert a strong
depressive influence upon these new contours and their for-
mation is delayed. What is assumed to take place is this:
when the bars disappear, their correSponding contours start
vanishing. The activity of off fibers dies out and on
fibers start firing. At first, some Strong interaction
between these two types of elements would take place.
Because the withdrawal of illumination corresponding to the
bars was gradient-like, this interaction would take place in
the retinal area covered by the bars and also would Spread
over several microns on either side, covering the inter-
Space between them. The line contours, having to form in
that Specific region would take more time to reach comple—
tion. Evidently, the exact mechanism remains to be worked
out. Whatever be this mechanism, however, the results, in
the case of the two observersfi clearly indicate that
destruction of contours markedly affect the formation of
new ones in their immediate surround, the two processes going
in opposite direction.

Shifting of border positions.——The results show that

 

less time is required for detection of the line when the
bars are shifted toward it in exposure period than when they
are Shifted away from it. When the bars, close together in

fore period, are farther away in exposure period, this is

 

 

 

77

what is assumed to take place: the bars contours completed
in fore period have to vanish in exposure period while

other contours corresponding to their new position are form-
ing. But the vanishing contours are close to the line con-
tours which are forming. It has been found, in Situation 4,
that this created much interference. It is worth mentioning
at this point that for the two observers the results obtained
in the present condition and in Situation 4 are comparable.
Evidently the main role was played by those contours which
were decaying and not by those which were forming at some
distance and simultaneously with the line contours.

When the Shifting was toward the center in exposure
period, the contours of the bars, formed in pre-exposure,
began to decay, but being at a certain distance from the
center, they did not exert any serious interference. At the
same time, contours corresponding to their new position were
building up in close vicinity to the line contours. The
bars being wider, their contours were more accentuated than
those of the line and the interference was at the expense
of these latter which required more time to complete. This
interference, however, was not as marked as when the bars
shifted away, because the two gradients, forming together,
were in the same direction, while in the other condition,
they were in opposite direction: one was building up while
the other was decaying.

Second, because the bars and line contours are build—

ing up simultaneously, less interference should be expected

 

 

 

78

than if the bars contours had been formed previously. This
expectation receives only partial support. It is validated
in the case of J.Y.L. who required less time for seeing the
line when the bars were shifting toward it than when they
remained close to it in fore and exposure periods (Situation
1). In the case of J.V. it does not hold. When strong and
weak contours were forming together (the present Situation
as well as Situations 2 and 3) he consistently required more
time than when already formed contours interfered with new
ones. This consistent difference between the two observers
is puzzling and the conditions in which it manifested itself
should be further investigated.

Another phenomenon which, in the case of the second
observer, seems to have affected his results in Situation 5
is apparent movement. It appears as though if the line can-
not be perceived before any apparent movement is experienced,
then this one will play a major role in the outcome. In
the present Situation some apparent movement was experi—
enced by the two observers. For example, when the target
in which the bars were set at a mediun distance was replaced
for a period of 25 milliseconds or so by the one in which
the bars were close together, a kind of rapid movement
first toward the center and then toward the periphery was
observed, it was like two bars colliding and then swinging
back to their original position. This movement became

really annoying, though, only when the second target was

79

presented for as long as 65 milliseconds approximately,
this annoying effect becoming more serious as duration was
increased. At high photic intensity the effect was found
to be much less distracting. Also, fixation became a very
important factor: with proper fixation apparent movement
became less objectionable.

When the Situation was reversed, that is when the
target wherein the bars were close together was presented
first, the distracting effect of apparent movement seemed
less marked and one had to reach a longer duration, or
approximately 145 milliseconds, in order to be affected by
it.

One of the observers (J.Y.L.) could detect the line
before reaching any of the values at which apparent movement
was clearly manifested. He remained well below 65 milli-
seconds when the bars were shifting toward the center, and
also he remained much below 145 milliseconds in the other
condition.

As for J.V., in the two conditions his results are
above those needed for experiencing beta movement. Moreover
he often complained about some movement producing a strong
distracting effect.

A comparison of the outcome in the present Situation
with some of Werner's results (1935) is interesting: when
two targets were presented in rapid succession, he found that

the contour corresponding to the second one, at a certain

80

critical interval, might prevent the formation of the con—
tours of the first one. This critical interval was found to
be the one at which apparent movement was optimal.

So, it seems reasonable to assume that if the con-
ditions for producing apparent movement are met, visual
acuity will be Significantly decreased. In the case of J.Y.L..
such conditions would have probably be attained by the
Experimenter using a thinner line or a lower degree of photic
intensity, for example.

To summarize and conclude it may be said that visual
acuity is affected by the presence of borders in vicinity.

(1) Distance between borders is one of the factors
affecting detection of a fine line. When these borders are
near the line, more time is required for its detection than
when they are far out. At a certain distance, however, they
have a facilitatory effect upon visual acuity.

(2) The length of borders in near vicinity destroys
the facilitatory effect of an increase in length of the
line.

(3) Removal of adjacent borders at the same time as
the line is presented has a strong depressive effect upon
visual acuity.

(4) Equally, a shifting of borders away from the line
requires more time for its detection.

(5) A Shifting of borders toward the line is facili-
tatory in the case of one of the observers, and inhibitory

in the case of the second one.

81

(6) Also, for one of the observers simultaneous pre-
sentation of the two adjacent borders and the line has a
facilitatory influence upon visual acuity, while the effect
is inhibitory for the other.

The results are discussed in terms of contour for—
mation. They suggest that any factor which hinders visual
acuity does so by interfering with this neural process.
Among the factors found to play a significant role are the
presence of other contours in vicinity, their order of

formation, and their destruction.

SUMMARY

This investigation examines the effects of adjacent
borders upon detection of a fine line.

A review of the literature on contour processes indi-
cates that borders near one another exert a mutually
depressive influence. This influence was Shown to extend
to distances as far as four degrees on the retina. If in a
set-up traditionally used for studying visual acuity bars
are introduced at different distances from a fine line to
be detected, one would expect visual acuity to be impaired
when these bars are in close vicinity of the line. As the
distance is increased the depressing effect of these bars
should progressively decrease. As far as the present author
knows, no study has been undertaken to investigate this.

Second, such a depressive effect, if present, can
be better understood by performing temporal manipulations
of borders. Thus, varying the order of presentation of the
bars and the line Should reveal Significant interactions
between contours. It Should also throw some light on the
relationship between contour formation and visual acuity.
Such manipulations have not been tried, however.

It was the purpose of this author to examine those
points, or more generally to analyze some of the implica-

tions of relating visual acuity and contour processes.

82

 

83

In the present investigation two trained observers

were tested for visual acuity under three different levels

of photic intensity in situations involving Spatial and

temporal manipulations of borders.

The apparatus was a modified Gerbrands Tachistoscope.

The material consisted of transparent cardboards which held

the fine line to be detected and the dark bars assumed to

exert an influence upon its detection.

The experiment comprised two parts. In Part One six

situations were investigated:

(1)

(2)

(3)

(4)

(5)

(6)

The distance between the two dark bars and the
line was decreased in five steps from 140 14'
to 11'.

The fore period was dark; the line and bars
appeared together in eXposure period.

The fore period was lighted, but the vertical
bars appeared simultaneously with the line in
exposure period.

The two bars vanished while the line appeared
in exposure period.

The two bars were Shifted toward the center or
toward the periphery.

The length of the line and of the two bars was

varied in three steps from 36' to 90 31'.

Part Two was performed as a check on the reliability

of data in those Situations where the two observers dis—

agreed. Thus Situations 2, 3, and 5 were re-examined.

 

84

The ascending method of limits was utilized through-
out the whole experiment.

The results indicate that visual acuity is affected
by the presence of borders in vicinity. (1) Distance be-
tween borders turned out to be a significant factor. When
these borders were near the line, more time was required
Lor its detection than when they were far out. At an inter—
mediary distance, howeyer, they had a facilitatory effect
upon visual acuity. (2) The length of borders in near
vicinity counterbalanced the facilitatory effect of an in-
crease in length of the line. (3) Removal of adjacent
borders at the same time as the line was projected has a
strong depressive effect upon visual acuity. (4) A shifting
of borders away from the line also required more time for
its detection. (5) A shifting of borders toward the line
was found to be facilitatory in the case of one observer,
and inhibitory in the case of the second one. (6) Equally,
for one observer simultaneous presentation of the two
adjacent borders and the line had a facilitatory influence
upon visual acuity, while the effect was inhibitory for the
other.

The results were discussed in terms of contour for—
mation. They suggest that any factor which impairs visual
acuity does so by interfering with this neural process.
Amont the factors found to play a significant role are the
presence of other contours in vicinity, their order of

formation, and their destruction.

BIBLIOGRAPHY

Arnulf, A., and Dupuy, 0. Contribution a l'étude des micro-
fluctuations d'accommodation de l'oeil. Rev. opt.,
1961, 39, 195-208.

Baker, K. E. Some variables influencing vernier acuity.
I- Illumination and exposure time. II— Wave—length
of illumination. J. opt. Soc. Amer., 1949, 39, 567-
576.

Bartley, S. H. The basis of the flicker in the visual field
surrounding the test-object. J. exp. Psychol., 1936,
3:2: 342‘550.

 

Bartley, S. H. Vision: A Study of its basis. New York:
Van Nostrand, 1941.

Bartley, S. H. Principles of perception. New York:
Harper, 1958 (a).

Bartley, S. H. Some facts and concepts regarding the neuro—
physiology of the optic pathway. A. M. A. Archives of
Ophthalmology, Part II, 1958, 69, 775-791 (b).

 

Bartley, S. H. Vision or sight. In Encyclopaedia britan-

 

nica, Vol. 23. Chicago: EncyclopaediafBritannica, 1960.

Bartley, S. H. A clarification of some of the procedures
and concepts involved in dealing with the optic path—
way. In R. Jung, and H. Kornhuber (Eds.), The visual
system: Neurgphysiology and psychophysics. Berlin:
Springer-Verlag, 1961.

 

Bartley, S. H., and Nelson, T. M. Some relations between
sensory end results and neural activity in the optic
pathway. J. Psychol., 1963, §§, 121-143.

Bartley, S. H., Nelson, T. M., and Soules, E. M. Visual
acuity under conditions of intermittent illumination
productive of paradoxical brightness. J. Psychol.,
1963,,§§, 153-163.

 

Blachowski, S. Ztschr. f. Sinnesphysiol., 1913, 41, Abt. II,
291. As reported in G. A. Fry and S. H. Bartley, The
effect of one border in the visual field upon the
threshold of another. Amer. J. Physiol., 1935, 112,

p. 421.

 

 

85

86

Boynton, R. M. Spatial vision. In Ann. Rev. Psychol.,
Palo Alto, Calif.: Annual Reviews Inc., 1962, 13,
171-200. ‘-

 

Campbell, F. W., and Westheimer, G. Dynamics of accommo-
dation responses of the human eye. JLgPhysiol.
(London), 1960, 151, 285-295.

Falk, J. L. Theories of visual acuity and their physio-
logical bases. Psychol. Bull., 1956, 53, 109-133.

 

Flom, M. C., Weymouth, F. W., and Kahneman, D. Visual
resolution and contour interaction. J. th. Soc.
Amer., 1963, 53, 1026-1032.

Fry, G. A., and Bartley, S. H. The effect of one border
in the visual field upon the threshold of another.
Amer. J. Physiol., 1935, 112, 414-421.

Fry, G. A., and Cobb, P. W. A new method for determining
the blurredness of the retinal image. Trans. Amer.
Acad. Ophthalmol. Otolaryngol., 1935, 49, 423-438.

Graham, C. H., and Cook, C. Minor studies from the Psycho—
logical Laboratory of Clark University. XXXII.
Visual acuity as a function of intensity and exposure-
time. Amer. J. Psychol., 1937, 49, 654-661.

Granit, R., and Harper, P. Comparative studies on the peri-
pheral and central retina. II. Synoptic reactions
in the eye. Amer. J. Physiol., 1930, 95, 211-228.

Hecht, S., and Mintz, E. U. The visibility of single lines
at various illuminations and the retinal basis of
visual resolution. J. gen. Physiol., 1939, 22, 593-

612.

Hecht, 8., Ross, 5., and Mueller, C. G. The visibility of
lines and squares at high brightnesses. J. opt. Soc.

Amer., 1947, g1, 500-507.

Helson, H., and Fehrer, E. V. The role of form in per-
ception. Amer. J. Psychol., 1932, 11, 79-102.

Jung, R. Neuronal integration in the visual cortex and
its Significance for visual information. In W. A.
Rosenblith (Ed.) Sensory communication. New York:

Wiley, 1961.

Kravkov, S. V. Illumination and visual acuity. Acta

 

87

Leibowitz, H. Some factors influencing the variability of
vernier adjustments. Amer. J. Psychol., 1955, 68,
266-273. . ‘—

Milner, P. M. Sensory transmission mechanisms. Canad. J.
Psychol., 1958, 1;, 149-158.

Nelson, T. M., Bartley, S. H., and De Hardt, D. A compari-
son of variability of three sorts of observers in a
sensory experiment. J. Psychol., 1960, 11, 3-11.

Nelson, T. M., Bartley, S. H., and Jewell, R. M. Effects
upon brightness produced by varying the length of the
null interval separating successive "single" pulses:
Sensory implications of the Alternation of ReSponse
Theory. J. Psychol., 1963, 56, 99-106.

Nelson, T. M., Bartley, S. H., and Wise, R. F. Size dis-
crimination under two conditions of photic inter—
mittency. J. Psychol., 1963, 56, 219-225.

Niven, J. R., and Brown, R. H. Visual resolution as a
function of intensity and exposure time in the human
fovea. J. opt. Soc. Amer., 1944, 31, 738-743.

Ogilvie, J. C., and Taylor, M. M. Effect of length on the
visibility of a fine line. J. Opt. Soc. Amer., 1959,
.12, 898-900.

Ogle, K. N. Foveal contrast thresholds with blurring of
the retinal image and increasing size of test stimulus.
J. opt. Soc. Amer., 1961, 51, 862-869.

Oliva, J., and Aguilar, M. Acuité visuelle et unites
sensorielles. Optica Acta. (Paris), 1957, 1, 145-147.

 

Osgood, C. E. Method and theory in experimental psychology.
New York: Oxford University Press, 1963.

Osterberg, G. Topography of the layer of rods and cones in
the human retina. Acta Ophthal. Supp_,, 1935, fii,

1-102 0

Ratliff, F. Inhibitory interaction and the detection and
enhancement of contours. In W. A. Rosenblith (Ed.),
Sensory communication. New York: Wiley, 1961.

Senders, V. L. The physiological basis of visual acuity.
Psychol. Bull., 1948, 15, 465-490.

Shlaer, S. The relation between visual acuity and illumi-
nation. J. gen. Physiol., 1937, 21, 165-188.

 

88

Werner, H. ~Studies on contour: I. Qualitative analyses.
Amer. J. Psychol., 1935, 11, 40-64.

Westheimer, G. Optical and motor factors in the formation
of the retinal image. J. opt. Soc. Amer., 1963, 53,
86-93.

Westheimer, G. Visual acuity. In Ann. Rev. Psychol.,
Palo Alto, Calif.: Annual Reviews Inc., 1965, 16,
359-380.

 

Westheimer, G., and Campbell, F. W. Light distribution in
the image formed by the living human eye. J. opt. Soc.
Amer., 1962, 51, 1040-1045.

 

Wilcox, W. W. The basis of the dependence of visual acuity
on illumination. Proc. nat. Acad. Sci., 1932,_1§,
47-56.

Wilcox, W. W. An interpretation of the relation between
visual acuity and light intensity. J. gen. Psychol.,
1936, 1§, 405-435.

 

RS

”'TIIIIIIIIIIIIIIIIIIIIIII’IIIIIIIIII'IIIIIIIIIIII“